Mr. J. Miers on Diclidanthera. 131 



calyx; five alternate petals spring from the margin of thecupnlar 

 disk, which is adnate to this cup; and ten anthers in one regular 

 scries are disposed on the summit of the adnate tube formed of 

 the confluent filaments, five being alternate and five opposite 

 to the petals, so that they thus appear sessile in the mouth of 

 the seemingly gamopetalous tube of the corolla. The structure 

 of these anthers offers a striking peculiarity : they are reniform, 

 compressed, erect, affixed by their basal sinus upon a very short 

 and broad portion of the filaments, which are conjoined below 

 into a monadelphous adnate tube, as before mentioned : these 

 anthers consist of two collateral cells, united into a compact 

 body ; and they open a little obliquely by a lateral fissure, or 

 rather by a hippocrepial line round their margin, into two un- 

 equal valves, the posterior retaining its naturally erect position, 

 the anterior valve becoming bent downwards. If we compare 

 this structure with the stamens of Biittneria, we perceive a close 

 similarity between them, — the chief difference being that in the 

 latter case the monadelphous tube of the stamens is free from 

 the corolla down to its base, the union of the filaments is there 

 not complete, and the free portions of the filaments appear as so 

 many segments of the border of the tube. Gay, in his mono- 

 graph of the family of the Biittneriacece, shows that in Commer- 

 sonia Fraseri* five of these stamens are anantherous, much 

 extended, and 3-fid, and five alternately short and antheriferous ; 

 its anthers are reniform, like those of Diclidanthera, and are 

 fixed by their sinus to the apex of the short broad filament, 

 which is reflexed at its apex, and the stamens appear as if they 

 were extrorse ; each anther, in like manner, opens by a hippo- 

 crepial line bilabiately, the internal valve remaining pendent, 

 and the more external valve rising into a vertical position, so 

 that the anther thus expanded is peltate, its cells being colla- 

 terally placed. If we imagine the monadelphous tube in this 

 case, as well as its free segments, to be agglutinated to the 

 corolla, as in Diclidanthera, the anthers would necessarily be 

 erect, instead of reflexed, we should find them placed exactly as 

 in that genus, and their dehiscence would be precisely similar. 



The ovary is globular, decidedly stipitate, as in Biittneria, 

 and in like manner five-celled, with two ovules, one of them 

 often rudimentary, suspended from the central column below 

 the apex of each cell by a very short funicle ; the priraine has 

 a wide tubular mouth in its summit, close to the funicle ; and the 

 raphe is distinguishable, extending along the ventral side to the 

 base. Although, in the Biittneria, the ovules are generally nu- 

 merous in each cell, they are reduced to two, placed collaterally, 

 in Rulingia. In the ripe fruit of Diclidanthera, each suspended 

 * Mem. du Mus. x. p. 215. tab. 15. figs. 8, 9, 10, & 11. 



9* 



